Housing Discrimination, Residential Racial Segregation, and Colorectal Cancer Survival in Southeastern Wisconsin Yuhong Zhou, Amin Bemanian, and Kirsten M.M

Home , Housing discrimination, Mortgage discrimination, Racial segregation

Published OnlineFirst February 14, 2017; DOI: 10.1158/1055-9965.EPI-16-0929 CEBP FOCUS: Geospatial Approaches to Cancer Control and Population Sciences

Housing Discrimination, Residential Racial Segregation, and Colorectal Cancer Survival in Southeastern Wisconsin Yuhong Zhou, Amin Bemanian, and Kirsten M.M. Beyer

Abstract

Background: Residential racial segregation is still neglected in blacks [HR ¼ 1.37; 95% confidence interval (CI), 1.06–1.76] and contemporary examinations of racial health disparities, including among black women (HR ¼ 1.53; 95% CI, 1.06–2.21), but not studies of cancer. Even fewer studies examine the processes by black men in sex-specific models. No associations were identified which segregation occurs, such as through housing discrimina- for redlining or the location quotient. Additional work is needed tion. This study aims to examine relationships among housing to determine whether these findings can be replicated in other discrimination, segregation, and colorectal cancer survival in geographical settings. southeastern Wisconsin. Conclusions: Our findings indicate that black women in par- Methods: Cancer incidence data were obtained from the Wis- ticular experience poorer colorectal cancer survival in neighbor- consin Cancer Reporting System for two southeastern Wisconsin hoods characterized by racial bias in mortgage lending, a measure metropolitan areas. Two indices of mortgage discrimination were of institutional racism. These findings are in line with previous derived from Home Mortgage Disclosure Act data, and a measure studies of breast cancer survival. of segregation (the location quotient) was calculated from U.S. Impact: Housing discrimination and institutional racism may census data; all predictors were specified at the ZIP Code Tabu- be important targets for policy change to reduce health dispa- lation Area level. Cox proportional hazards regression was used to rities, including cancer disparities. Cancer Epidemiol Biomarkers Prev; examine associations between mortgage discrimination, segrega- 26(4); 1–8. 2017 AACR. tion, and colorectal cancer survival in southeastern Wisconsin. See all the articles in this CEBP Focus section, "Geo- Results: For all-cause mortality, racial bias in mortgage lending spatial Approaches to Cancer Control and Population was significantly associated with a greater hazard rate among Sciences."

Introduction Researchers are beginning to explore additional factors that may contribute to racial disparities in cancer outcomes, including Colorectal cancer is the third leading cause of cancer-related residential racial segregation (13–15). Research on breast cancer death in both men and women in the United States (1) and has hypothesized linkages between segregation and survival colorectal cancer survival disparities, including by race and geog- through health care access, exposure to stressors, and local health raphy, have been extensively documented (2, 3). Despite the behavioral norms, including physical activity, nutrition, tobacco, continuing decrease in colorectal cancer death rates over the past and alcohol use (16, 17), and a few studies have indicated that two decades, racial/ethnic minority groups, particularly blacks/ segregation may contribute to racial disparities in cancer mortal- African Americans, continue to have higher death rates compared ity, although findings have been mixed (14, 17–19). There are far with whites (1, 4, 5). The gap in colorectal cancer survival rates by fewer publications investigating the effects of racial residential race has persisted since the early 1980s (2, 3) and may be growing segregation on colorectal cancer outcomes. In a study in the Twin wider (6–8). Studies have shown that diagnosis stage, tumor Cities 7-county Metropolitan area, Shen (20) found that facilities biology and genetics, comorbidities, lifestyle factors, differences that were located closer to minority-segregated census tracts had in screening and treatment, and socioeconomic status all can play poorer colorectal cancer screening performance. In a study explor- a role in generating racial disparities in colorectal cancer mortality ing the association of segregation with disparities in diagnosis and survival (9–12). However, even after controlling these known stage for breast, colorectal, prostate, and lung cancer, Haas and contributing factors, the racial survival gap for colorectal cancer is colleagues (21) found that the black/white disparity was actually not fully eliminated (9). smaller in more segregated areas, after adjusting for individual- level factors and an area-level urban/rural indicator. Given the plausibility of the influence of segregation on colorectal cancer Division of Epidemiology, Institute for Health and Equity, Medical College of survival (13, 22), additional work is needed to better understand Wisconsin, Milwaukee, Wisconsin. these relationships. Corresponding Author: Yuhong Zhou, Medical College of Wisconsin, 8701 Although studies of segregation are greatly needed, studying Watertown Plank Road, Milwaukee, WI 53226. Phone: 414-955-4302; Fax: only segregation patterns may not provide sufficient information 414-955-0176; E-mail: [email protected] to inform policy change to improve health and reduce disparities. doi: 10.1158/1055-9965.EPI-16-0929 Segregation measures reveal spatial distributions of population 2017 American Association for Cancer Research. groups by race, but do not directly measure the underlying

www.aacrjournals.org OF1

discriminatory and socioeconomic processes that create the pat- purpose of cancer prevention and control as defined in Statute terns. Processes contributing to segregation patterns involve mul- 255.04(3)(c). tiple sectors of society, such as housing, education, and labor (23), The HMDA database was initially created to collect data on and are ultimately the targets for policy changes to reduce resi- mortgage-lending practices. It reports relevant information on dential racial segregation. Recently, several studies have found mortgage applications, such as applicants' demographic and relationships between mortgage discrimination and health status, economic characteristics (race/ethnicity, sex, and income), prop- including pregnancy health, preterm birth, and most recently, erty type, loan purpose, loan amount, and mortgage decision. The breast cancer survival (24–27), indicating that racial discrimina- census tract containing the residential address of the property for tion in housing could be important in explaining racial/ethnic which a mortgage was requested is also included. Data were disparities in health outcomes, including cancer. No studies have limited to applications for purchasing an owner-occupied home looked at the relationship between patterns of housing discrim- and without missing information on the primary race/ethnicity, ination and colorectal cancer survival, and none have concurrent- sex and income of the primary applicant, loan amount, and ly examined housing discrimination and segregation. whether the loan was denied. Of a total of 396,032 total applica- The purpose of this study is to examine relationships between tions for the purchase of an owner-occupied home, approximately housing discrimination segregation, and colorectal cancer surviv- 40% of applications were missing data on at least one of these al, contributing to a growing body of work examining racism, variables; 32% of applications were missing approval/denial segregation, and cancer outcomes. status. To mitigate common problems with estimation at study area boundaries, estimates calculated near the boundaries of the Materials and Methods study area also included data from census tracts in counties outside of, but bordering, the study area. Study area The outcome variable is the survival time after colorectal The study area includes two metropolitan statistical areas cancer diagnosis, which is calculated as the number of months (Milwaukee-Waukesha-West Allis and Racine) in southeastern between initial diagnosis and either date of death or December Wisconsin. As the center of this region, the City of Milwaukee 31, 2011 (the last day of the study period). Two censoring is home to approximately 600,000 residents, of whom non- variables were used on the basis of cause of death information Hispanic black and non-Hispanic white populations share to reflect (i) colorectal cancer as the underlying cause of death similar percentages (39% and 38%). The population within and (ii) all causes of death among men and women diagnosed Milwaukee County and the City of Milwaukee experiences with colorectal cancer. The first variable reflects censoring of lower socioeconomic status (SES) than the state population, individuals who died of causes other than colorectal cancer, or including lower incomes, higher poverty, greater unemploy- were alive on the last day of the study period, whereas the ment, lower educational attainment, lower home ownership second variable reflects censoring of only those alive on the last rates, and poorer housing stability (28). The long-term day of the study period. entrenched poverty and residential racial segregation in Mil- Primary predictors included two new indices of mortgage waukee, its history of discriminatory housing policies (29), and discrimination and one segregation metric. Following the work observed disparities in colorectal cancer incidence and mortal- of Beyer and colleagues (27), we calculated two indices, racial bias ity rates (30) make the area an appropriate setting for this in mortgage lending and residential redlining, to measure hous- study. Figure 1 displays the geographic extent of the study area, ing discrimination. Both indices were estimated by integrating as well as patterns of colorectal cancer incidence and mortality logistic regression models with the adaptive spatial filtering (ASF) in the region, to provide context for the study. approach (33, 34). To apply ASF, a grid is first laid over the study area, and spatial filters symbolized by circles then are created and centered at each grid point. The idea of ASF is to expand the radius Data and variables of the filter for each grid point until enough observations from Our analyses are based on three data sources. Cancer incidence nearby geographic units (census tracts in this case) falling within data were provided by the Wisconsin Cancer Reporting System for the filter are obtained to calculate a stable statistic. The statistic is the years 2002 to 2011 for invasive colorectal cancers for south- mapped as a continuous surface using inverse distance weighted eastern Wisconsin. Segregation metrics were calculated from U.S. method. The racial bias in mortgage lending index is the statistic Census Bureau population and demographic data (31). Indices of estimated for each grid point using the observations within the mortgage discrimination were derived from Home Mortgage Dis- filter. It is the odds of denial of a mortgage application from a closure Act (HMDA) data (2004–2011) available on the Federal black primary applicant compared with denial of a white primary Financial Institutions Examination Council HMDA website (32). applicant, while controlling for sex, and the ratio of the loan Reported by hospitals, physicians, and clinics directly to Wis- amount to the applicant's gross annual income. We used a consin Department of Health Services (DHS), cancer cases include threshold of a minimum of five denied black applicants and five important information about patients' demographics, tumor denied white applicants to guide the filter size. In contrast, the characteristics and treatment, date and cause of death via linkages residential redlining index measures the bias against issuing to the Wisconsin Vital Records resident death file and National mortgages in particular neighborhoods. Thus, the redlining index Death Index. The sample used was limited to individuals who are is constructed by estimating the odds of denial of the mortgage black or non-Hispanic white and resided in the study area at application for individuals inside the filter, as compared with diagnosis. Cases missing diagnosis stage information were exclud- individuals outside the filter. The same filter threshold was ed (<3%). This study was approved by the institutional review applied to calculate this index. We derived two variations of the board at the local institution and authorized and approved by the redlining measure, one only adjusting for sex and loan amount to DHS Research Review Board for the release of cancer data for the income ratio and another controlling also the race and ethnicity of

OF2 Cancer Epidemiol Biomarkers Prev; 26(4) April 2017 Cancer Epidemiology, Biomarkers & Prevention

Housing Discrimination, Segregation, and Colorectal Cancer

Figure 1. Patterns of colorectal cancer incidence (A) and mortality (B) in the study area. The invasive colorectal cancer incidence/mortality rate is indirectly age-sex standardized and smoothed using the ASF approach. A grid of points is used to estimate incidence/mortality rates continuously across the map, based on the 30/20 closest diagnosed/colorectal cancer mortality cases. Darker areas indicate higher rates than expected and lighter areas indicate lower rates than expected, given the regional rate. Incidence data are from the Wisconsin Cancer Reporting System (WCRS), and mortality data are from the State Vital Records Office. the primary applicant. For inclusion in survival models, both Massey and Denton, quantify how groups are separated within indices represented by the interpolated continuous surfaces were a given area (36). Therefore, they are not suitable for measuring summarized (mean pixel value) by ZIP Code Tabulation Area the effect of small area's segregation on a larger region. The unit (ZCTA). We also derived binary predictors from them. The binary we employed is ZCTA, to which the colorectal cancer cases were variable of racial bias index is coded as 1 if the index value is equal geocoded, and the region is confinedtothestudyarea.LQ to 2 or greater and as 0 otherwise. The cutoff value for coding the ranges from zero to infinity. An LQ equal to zero indicates that binary variable of the redlining index is 1. Although the two there are no residents of group m in the neighborhood unit, redlining measures were both tested, we only reported the results whereas an LQ less than one indicates that the proportion of for the one that is race/ethnicity adjusted, as results were similar. group m in the neighborhood is less than the proportion of the Segregation was measured at the ZCTA level using the location same group in the region. On the basis of the work of Pruitt and quotient (LQ), a measure of local area segregation (19, 35). The colleagues (19), the LQ was log(xþ1) normalized. Calculation equation for calculating LQ is as follows: of the LQ and mortgage discrimination metrics was completed in R (37) and ArcGIS (38). xim=Xi LQim ¼ Xm=X Statistical analysis We used multivariable Cox proportional hazards regression to model survival time for both colorectal cancer–specific mortality Where LQim is the value for the ith unit in a region for and all-cause mortality among black and non-Hispanic white population group m (black in our case), xim is the number of individuals diagnosed with incident colorectal cancer in the study individuals of the mth group living in the ith unit, Xi is the total area between 2002 and 2011. In addition to the LQ for the black number of residents in the ith unit of the region, Xm is the total population and two new indices of mortgage discrimination number of individuals from minority group m in the region, (each incorporated into the survival model, one at a time), and X is the total number of residents living in the region. The individual characteristics, such as age (18–44, 45–54, 55–64, LQ relates the proportion of individuals in a local area of a 65–74, and 75þ), sex (male and female), and stage at diagnosis particular race to the same proportion at the regional level. (SEER Summary Stage 2000 categories, local, regional, and dis- Conceptually, the LQ represents the relative concentration of a tant), were included as control variables. Models also control for racial group and can explain how the demographic makeup of a two neighborhood-level variables, ZCTA population density and single small unit contributes to the overall racial distribution in an index of ZCTA socioeconomic status. The index was estimated the metropolitan area (35). The traditional measures of segre- using principle component analysis for selected American Com- gation, such as dissimilarity and exposure, as outlined by munity Survey variables: median household income, percent

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 26(4) April 2017 OF3

unemployed, percent renter households, percent families led by measures of segregation or mortgage discrimination were asso- single female, and percent poverty. For the continuous redlining ciated with colorectal cancer survival. index and LQ, two different models with/without adding population density as a covariate were tested due to a concern about the Discussion moderate correlation between population density and the two primary predictors. The results for modeling the effects of con- This study contributes new knowledge to a small but growing tinuous redlining index with/without controlling population body of research regarding institutional racism, segregation, and density are similar in terms of direction, magnitude, and signif- cancer outcomes, helping to shed light on possible directions for icance; thus, we only reported the model results with inclusion of policy change and public health intervention. This is the first study population density. In total, there are nine models being tested for to examine linkages between elements of mortgage discrimina- each outcome variable (six of which are reported in the results tion and segregation concurrently, in their association with cancer section), examining the LQ and continuous and binary versions of survival, and the first to examine the relationship between mort- the racial bias index and two redlining indices. Survival analyses gage discrimination and colorectal cancer survival. We found that were implemented in Stata SE/13 (39) and R (37).The propor- racial bias in mortgage lending (when measured as a binary tional hazards assumption was examined for the models with all variable) was related to poorer colorectal cancer survival among the predictors (age, sex, stage at diagnosis, population density, blacks, but not among whites, and that this association was driven and segregation or mortgage discrimination index), and it was by a strong relationship for black women. Neither redlining nor found that the stage variable often violated the assumption. Thus, the LQ exhibited any statistically significant associations, and the we applied stratified Cox model to correct the problem, using the only variable of importance for white patients was population diagnosis stage as a stratification variable. In addition, we fitted density, with higher density associated with a higher hazard rate. models with frailty terms for ZCTAs to examine the possibility of The measures employed in this study are conceptually dis- spatial clustering, but no frailty terms were statistically significant. tinct. Whereas the index of racial bias in mortgage lending Finally, we fitted additional, sex-specific models for the black indicates the odds of denial of a mortgage application in a population to determine whether observed relationships held for particular area by race, the redlining index measures the odds of both sexes. denial of a mortgage application in a particular neighborhood, regardless of race. The LQ seeks to relate the proportion of a racial group in the local area to the same proportion in the Results larger region, providing a sense of the relative degree of seg- Figure 2 displays the spatial distributions of the racial bias regation in a specific local area. In the study area, the LQ and index, redlining index while controlling for the race and ethnicity redlining indices reveal more similar spatial patterns, with of the primary applicant, and black LQ. higher values in Milwaukee's predominantly black central city. Table 1 presents descriptive statistics for the population under In contrast, the racial bias index tends to be higher in areas study. All individuals in the sample are black/African American or outside of the central city, where fewer black residents reside, non-Hispanic white individuals diagnosed with colorectal cancer reinforcing their status as a racial minority. between 2002 and 2011 in the study area. The sex composition is There are a number of possible explanations for our finding that approximately half females and half males. A relatively large institutional racism is associated with poorer survival after colo- proportion of the individuals in the sample was diagnosed with rectal cancer diagnosis. The long-term presence of race-related a localized tumor (39.59% among blacks and 41.16% among mortgage discrimination in the Milwaukee area, as a manifesta- whites). Of those deceased (265 blacks and 1,940 whites), tion of institutional racism, could be a persistent source of stress or 67.17% and 59.74% died from colorectal cancer. a barrier to health care access or utilization, thus promoting Tables 2–4 show the results of Cox proportional hazards progression of colorectal cancer or hindering recovery and leading models for the racial bias index, redlining, and the location to shorter survival of black colorectal cancer patients. The higher quotient, respectively. For all-cause mortality, the binary racial likelihood of black populations being denied to access to financ- bias in mortgage lending variable was significantly associated with ing for housing could indicate possible discrimination in other a greater hazard rate for blacks [HR ¼ 1.37; 95% confidence sectors, reducing access to other resources important to their interval (CI), 1.06–1.76 in Model 1.2, Table 2], but not for whites. health and medical needs after being diagnosed with colorectal The racial bias index was not significantly related to colorectal cancer. However, it is unclear why this relationship would affect cancer–specific mortality. The redlining index and LQ (Tables 3 black women but not black men. The small sample size does not and 4) were not significantly associated with all-cause mortality or appear to have played a role in negative findings for males, as a colorectal cancer–specific mortality. Of note, although it was not a higher proportion of colorectal cancer patients in the database primary predictor of interest, higher population density was were male. The sex-specific aspects of these relationships require significantly and consistently associated with a higher hazard rate further study. among whites. Interestingly, results did not indicate that individuals living in Table 5 shows the results of Cox proportional hazards models redlined neighborhoods, or those characterized by high levels of for all-cause mortality among black women diagnosed with local segregation, experienced poorer colorectal cancer–specific colorectal cancer. Racial bias in mortgage lending (binary) was survival. These findings may seem to run counter to intuition, but associated with poorer colorectal cancer survival among black do reflect findings from some other studies (14, 27). It is possible women, with an HR comparable with but larger than that that black patients could be exposed to protective factors in observed for the full sample (HR ¼ 1.53; 95% CI, 1.06–2.21 in predominantly black central city neighborhoods. In particular, Model 4.1.2, Table 5). Neither the redlining index nor the LQ was the presence of strong social networks and social support may significantly associated with survival. In male-only models, no mitigate the effects of discrimination or generate protective effects

OF4 Cancer Epidemiol Biomarkers Prev; 26(4) April 2017 Cancer Epidemiology, Biomarkers & Prevention

Housing Discrimination, Segregation, and Colorectal Cancer

Figure 2. The racial bias in mortgage lending index maps (A and B in the top row), the race- and ethnicity-adjusted redlining index maps (D and E in the bottom row), the black LQ map (C in the top row). The graphs (A, B, D,andE) for mortgage discrimination measures represent the time period from 2004 to 2011 and are based on tract-level HMDA data. Each of these graphs in a row presents the average pixel value for the index for each ZCTA and the study area divided into two categories as described for binary indices.

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 26(4) April 2017 OF5

Table 1. Sample characteristics another [of coapplicant(s)] when deriving indices has not yet White (n ¼ 4,699) Black (n ¼ 682) been explored. It would be useful to examine the change in Frequency (%) Frequency (%) patterns of mortgage discrimination when using different Sex definitions of race and ethnicity. Third, the measures of seg- Male 2,465 (52.46%) 326 (52.20%) Female 2,234 (47.54%) 356 (47.80%) regation and mortgage discrimination we used only capture the Age group static condition of segregation and mortgage denial patterns 18–44 years 254 (5.40%) 66 (9.68%) during a specific time period, while colorectal cancer patients 45–54 years 663 (14.11%) 170 (24.93%) were diagnosed at different times and their exposures to the 55–64 years 865 (18.41%) 184 (26.98%) adverse effects of segregation and discrimination may change – 65 74 years 1,005 (21.39%) 150 (21.99%) with their mobility and life experience. Finally, as measure- 75þ years 1,912 (40.69%) 112 (16.42%) SEER Summary Stage 2000 categories ment is imperfect and not all important factors can be mea- Localized 1,934 (41.16%) 270 (39.59%) sured with available data, there is of course the possibility that Regional 1,882 (40.05%) 247 (36.22%) residual or unmeasured confounding could impact effect esti- Distant 883 (18.79%) 165 (24.19%) mates for the exposures of interest. However, we did control for Vital status (as of December 31, 2011) SES, a major potential source of unmeasured confounding. Alive 2,759 (58.71%) 417 (61.14%) Furthermore, we did not find evidence of spatial clustering Deceased 1,940 (41.29%) 265 (38.86%) Cause of death (n ¼ 1,940) (n ¼ 265) using frailties, providing little evidence of unmeasured spa- Colorectal cancer 1159 (59.74%) 178 (67.17%) tially varying confounders. Other causes 781 (40.27%) 87 (32.83%) Despite these limitations, this study presents a novel perspec- tive on the role of housing discrimination and segregation in racial disparities in colorectal cancer survival. Future work should (40). Furthermore, institutional racism in health care access examine whether such findings hold for other racial and ethnic experienced by black patients living in predominantly white areas groups, and in other geographic settings. In addition, more may be less of a barrier in areas with higher black populations, as research is needed to elucidate the pathways by which segregation encounters between black patients and physicians are likely to be influences cancer survival disparities and to move these findings more common. Further study is needed to untangle the practical toward intervention and population health improvement. In mechanisms affecting disparate populations. addition, there are other social/economic processes that contrib- Our work has several limitations. First, the HMDA data do ute to segregation patterns and are worth exploring, such as not include several variables that may affect denial rates, discrimination in education and labor. Finally, it would be including current employment and credit scores. Second, there interesting to explore survival between colon and rectum cancer, are multiple variables for race and ethnicity in the HMDA instead of combining them as colorectal cancer, as their risks for database and the sensitivity of modeling results to the choice of men and women are different (41, 42). More translational one variable (race/ethnicity of the primary applicant) versus research on policy development and intervention practices are

Table 2. Cox proportional hazards regression models relating racial bias in mortgage lending to all-cause and colorectal cancer–speciﬁc mortality Black patients (n ¼ 682) White patients (n ¼ 4,699) Model 1.1 Model 1.2 Model 1.3 Model 1.4 HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) All-cause survival Population density (100 per sq km) 0.88 (0.64–1.23) 0.88 (0.64–1.21) 1.23a (1.06–1.44) 1.27a (1.08–1.48) Racial bias index (continuous) 1.02 (0.92–1.13) — 1.01 (0.98–1.04) — Racial bias index (binary; 2) — 1.37a (1.06–1.76) 0.95 (0.84–1.08) Colorectal cancer–speciﬁc survival Population density (100 per sq km) 0.93 (0.62–1.39) 0.92 (0.62–1.37) 1.24a (1.02–1.51) 1.27a (1.04–1.55) Racial bias index (continuous) 1.02 (0.91–1.14) — 1.00 (0.97–1.04) — Racial bias index (Binary; 2) — 1.25 (0.92–1.70) 0.95 (0.81–1.12) aP < 0.05

Table 3. Cox proportional hazards regression models relating redlining to all-cause and colorectal cancer–speciﬁc mortality Black patients (n ¼ 682) White patients (n ¼ 4,699) Model 2.1 Model 2.2 Model 2.3 Model 2.4 HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) All-cause survival Population density (100 per sq km) 1.05 (0.73–1.52) 0.89 (0.64–1.25) 1.24a (1.07–1.44) 1.23a (1.06–1.43) Redlining index (continuous) 0.82 (0.66–1.02) — 1.01 (0.85–1.20) — Redlining index (binary; 1) — 1.08 (0.63–1.86) — 0.96 (0.84–1.09) Colorectal cancer–speciﬁc survival Population density (100 per sq km) 1.11 (0.71–1.73) 0.89 (0.59–1.35) 1.25a (1.03–1.51) 1.23a (1.06–1.51) Redlining index (continuous) 0.80 (0.62–1.05) — 1.05 (0.84–1.31) — Redlining index (binary; 1) — 0.70 (0.38–1.29) — 0.96 (0.81–1.15) aP < 0.05

OF6 Cancer Epidemiol Biomarkers Prev; 26(4) April 2017 Cancer Epidemiology, Biomarkers & Prevention

Housing Discrimination, Segregation, and Colorectal Cancer

Table 4. Cox proportional hazards regression models relating local segregation (Black LQ) to all-cause and colorectal cancer–speciﬁc mortality Black patients (n ¼ 682) White patients (n ¼ 4,699) Model 3.1 Model 3.2 Model 3.3 Model 3.4 HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) All-cause survival Population density (100 per sq km) 0.89 (0.64–1.24) — 1.29a (1.10–1.52) — Black LQ 0.62 (0.29–1.36) 0.62 (0.29–1.34) 1.28 (0.89–1.84) 1.03 (0.73–1.44) Colorectal cancer–speciﬁc survival Population density (100 per sq km) 1.08 (0.62–1.40) — 1.26a (1.02–1.56) — Black LQ 0.75 (0.29–1.98) 0.75 (0.29–1.97) 1.09 (0.67–1.75) 0.88 (0.57–1.36) aP < 0.05

Table 5. Cox proportional hazards regression models predicting all-cause mortality among black women diagnosed with colorectal cancer (n ¼ 356) Model 4.1.1 Model 4.1.2 Model 4.2.1 Model 4.2.2 Model 4.3.1 Model 4.3.2 HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) HR (95% CI) Population density (100 per sq km) 0.67 (0.40–1.34) 0.68 (0.42–1.10) 0.77 (0.45–1.31) 0.64 (0.37–1.11) 0.68 (0.41–1.12) — Racial bias index (continuous) 1.12 (0.94–1.34) —— ——— Racial bias index (binary; 2) — 1.53a (1.06–2.21) ———— Redlining index (continuous) ——0.89 (0.66–1.20) ——— Redlining index (binary; 1) ———0.65 (0.29–1.41) —— Black LQ ————0.31 (0.09–1.16) 0.38 (0.11–1.32) aP < 0.05 needed to achieve the ultimate goal of reducing the impact of Administrative, technical, or material support (i.e., reporting or organizing institutional racism, including mortgage discrimination, on pop- data, constructing databases): Y. Zhou, K.M.M. Beyer ulation health. Study supervision: K.M.M. Beyer Other (postdoctoral mentor of the first author): K.M.M. Beyer Disclosure of Potential Conflicts of Interest No potential conflicts of interest were disclosed. Grant Support Authors' Contributions All authors received support from Research and Education Program Fund, a Conception and design: Y. Zhou, K.M.M. Beyer component of the Advancing a Healthier Wisconsin endowment at the Medical Development of methodology: K.M.M. Beyer College of Wisconsin and in part by the Medical College of Wisconsin Cancer Acquisition of data (provided animals, acquired and managed patients, Center, Population Sciences Program (principal investigator, K.M.M. Beyer). provided facilities, etc.): Y. Zhou, K.M.M. Beyer Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Zhou, A. Bemanian, K.M.M. Beyer Writing, review, and/or revision of the manuscript: Y. Zhou, A. Bemanian, Received November 15, 2016; revised January 9, 2017; accepted February 10, K.M.M. Beyer 2017; published OnlineFirst February 14, 2017.

References 1. American Cancer Society. Cancer facts and figures 2016. Atlanta, GA: 10. Polite BN, Dignam JJ, Olopade OI. Colorectal cancer model of health American Cancer Society; 2016. disparities: understanding mortality differences in minority populations. J 2. Doubeni CA, Rustgi A. Racial disparities in colorectal cancer survival: Is Clin Oncol 2006;24:2179–87. elimination of variation in care the cure? J Natl Cancer Inst 2015;107: 11. Le H, Ziogas A, Lipkin SM, Zell JA. Effects of socioeconomic status and djv229. treatment disparities in colorectal cancer survival. Cancer Epidemiol Bio- 3. Ward E, Jemal A, Cokkinides V, Singh GK, Cardinez C, Ghafoor A, et al. markers Prev 2008;17:1950–62. Cancer disparities by race/ethnicity and socioeconomic status. CA Cancer J 12. White A, Vernon SW, Franzini L, Du XL. Racial disparities in colorectal Clin 2004;54:78–93. cancer survival: to what extent are racial disparities explained by differences 4. American Cancer Society. Cancer facts and figures for African Americans in treatment, tumor characteristics, or hospital characteristics? Cancer 2016-2018. Atlanta, GA: American Cancer Society; 2016. 2010;116:4622–31. 5. Howlader N, Noone A, Krapcho M, Miller D, Bishop K, Altekruse S, et al. 13. Williams DR, Collins C. Racial residential segregation: a fundamental cause SEER Cancer Statistics Review, 1975-2013. Bethesda, MD: NIH; 2016. of racial disparities in health. Public Health Rep 2001;116:404–16. 6. Hunt B, Whitman S. Black:White health disparities in the United States and 14. Warner ET, Gomez SL. Impact of neighborhood racial composition and Chicago: 1990–2010. J Racial Ethn Health Disparities 2015;2:93–100. metropolitan residential segregation on disparities in breast cancer stage at 7. Robbins AS, Siegel RL, Jemal A. Racial disparities in stage-specific colorectal diagnosis and survival between black and White women in California. J cancer mortality rates from 1985 to 2008. J Clin Oncol 2012;30:401–5. Community Health 2010;35:398–408. 8. Soneji S, Iyer SS, Armstrong K, Asch DA. Racial disparities in stage-specific 15. Broder S. Progress and challenges in the national cancer program. Plain- colorectal cancer mortality: 1960-2005.AmJ Public Health 2010;100:1912–6. view, NY: Cold Spring Harbor Laboratory Press; 1991. 9. Alexander DD, Waterbor J, Hughes T, Funkhouser E, Grizzle W, Manne U. 16. Russell E, Kramer MR, Cooper HLF, Thompson WW, Arriola KRJ. African-American and Caucasian disparities in colorectal cancer mortality Residential racial composition, spatial access to care, and breast cancer and survival by data source: an epidemiologic review. Cancer Biomarkers mortality among women in Georgia. J Urban Health 2011;88: 2007;3:301–13. 1117–29.

www.aacrjournals.org Cancer Epidemiol Biomarkers Prev; 26(4) April 2017 OF7

17. Schootman M, Jeffe DB, Gillanders WE, Aft R. Racial disparities in the 28. U.S. Census Bureau. Quickfacts 2015. Suitland, MD: U.S. Census Bureau; development of breast cancer metastases among older women: a multilevel 2015. study. Cancer 2009;115:731–40. 29. Gurda J, Looze C. The making of Milwaukee. Milwaukee, WI: Milwaukee 18. Pruitt SL, Lee SC, Ruiz JM, Xuan L, Tiro JA. Examining the Latina paradox: County Historical Society; 1999. impact of neighborhood Latina density on breast cancer survival in Texas. 30. Beyer KMM, Zhou Y, Matthews K, Hoormann K, Bemanian A, Laud PW, Cancer Epidemiol Biomarkers Prev 2015;24:PR12–PR12. et al. Breast and colorectal cancer survival disparities in Southeastern 19. Pruitt SL, Lee SJC, Tiro JA, Xuan L, Ruiz JM, Inrig S. Residential racial Wisconsin. WMJ 2016;115:17–21. segregation and mortality among black, white, and Hispanic urban breast 31. US Census Bureau. Census 2010. American FactFinder. Suitland, MD: U.S. cancer patients in Texas, 1995 to 2009. Cancer 2015;121:1845–55. Census Bureau; 2010. 20. Shen Q. Investigating the effects of racial residential segregation, area-level 32. Federal Financial Institutions Examination Council. Home Mortgage Dis- socioeconomic status and physician composition on colorectal cancer closure Act. 2017. Available from: https://www.fﬁec.gov/hmda/ screening [dissertation]. Richmond, VA: Virginia Commonwealth Univer- 33. Tiwari C, Rushton G. Using spatially adaptiveﬁlters to map late stage colorectal sity; 2016. cancer incidence in Iowa. In: Fisher PF, editor. Developments in Spatial Data 21. Haas JS, Earle CC, Orav JE, Brawarsky P, Neville BA, Williams DR. Racial Handling. Berlin/Heidelberg, Germany: Springer-Verlag; 2005; p. 665–76. segregation and disparities in cancer stage for seniors. J Gen Intern Med 34. Beyer KMM, Tiwari C, Rushton G. Five essential properties of disease maps. 2008;23:699–705. Ann Assoc Am Geogr 2012;102:1067–75. 22. Collins CA, Williams DR. Segregation and mortality: the deadly effects of 35. Sudano JJ, Perzynski A, Wong DW, Colabianchi N, Litaker D. Neighbor- racism? Sociol Forum 1999;14:495–523. hood racial residential segregation and changes in health or death among 23. Harrison KM, Dean HD. Use of data systems to address social determinants older adults. Health Place 2013;19:80–8. of health: a need to do more. Public Health Rep 2011;126:1–5. 36. Massey DS, Denton NA. The dimensions of residential segregation. Soc 24. Gee GC. A multilevel analysis of the relationship between institutional and Forces 1988;67:281–315. individual racial discrimination and health status. Am J Public Health 37. R Development Core Team. R: a language and environment for statistical 2002;92:615–23. computing. Vienna, Austria: R Development Core Team; 2011. 25. Mendez DD, Hogan VK, Culhane J. Institutional racism and pregnancy 38. ESRI. ArcGIS for Desktop. 2016. Available from: http://desktop.arcgis. health: using Home Mortgage Disclosure Act data to develop an index for com/en/ mortgage discrimination at the community level. Public Health Rep 39. StataCorp. Stata: Release 13. Statistical Software. 2013. Available from: 2011;126:102–14. https://www.stata.com/stata13/ 26. Mendez DD, Hogan VK, Culhane JF. Stress during pregnancy: the role of 40. Kramer MR, Hogue CR. Is segregation bad for your health? Epidemiol Rev institutional racism. Stress Health 2013;29:266–74. 2009;31:178–94. 27. Beyer KMM, Zhou Y, Matthews K, Bemanian A, Laud PW, Nattinger AB. 41. Lee Y-C, Lee Y-L, Chuang J-P, Lee J-C. Differences in survival between colon New spatially continuous indices of redlining and racial bias in mortgage and rectal cancer from SEER data. PLoS One 2013;8:e78709. lending: links to survival after breast cancer diagnosis and implications for 42. Edge S, Byrd D, Compton C, Fritz A, Greene F, Trotti A. AJCC cancer staging health disparities research. Health Place 2016;40:34–43. handbook. 7th ed. New York, NY: Springer; 2010.

OF8 Cancer Epidemiol Biomarkers Prev; 26(4) April 2017 Cancer Epidemiology, Biomarkers & Prevention

Housing Discrimination, Residential Racial Segregation, and Colorectal Cancer Survival in Southeastern Wisconsin

Yuhong Zhou, Amin Bemanian and Kirsten M.M. Beyer

Cancer Epidemiol Biomarkers Prev Published OnlineFirst February 14, 2017.

Updated version Access the most recent version of this article at: doi:10.1158/1055-9965.EPI-16-0929

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cebp.aacrjournals.org/content/early/2017/03/20/1055-9965.EPI-16-0929. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.